Story perspectives

U.S. banks block AI agents from customer accounts

By Drooid · Oct 28, 2026, 14:13 UTC · How we work

Read the story

Story summary

Summary

U.S. banks block third-party AI agents from logging into customer accounts.
Bank of America warned customers on the 26th not to share usernames, passwords, or AI tool access.
The Bank of England and FCA reported 79% of UK banks rely on the three largest cloud providers.
Anthropic warned in its IPO filing that AI liability could become enormous and insurers may be scarce.

Summary

Bank of America, Citi, Morgan Stanley and Blackstone pledged $63 billion to fund Broadcom AI hardware for Anthropic.
A special-purpose vehicle will purchase the equipment and lease it to Anthropic.
First-lien senior debt amounts to about $43.5 billion, roughly 69 % of the total commitment.
Broadcom offers residual-value support to protect lenders against depreciation of the AI accelerators.
Moody’s warned that the rapid AI-related financings could raise Broadcom’s risk profile.

History

The 2010 Dodd-Frank Wall Street Reform and Consumer Protection Act, enacted after the 2008 financial crisis, imposed sweeping oversight on banks. The Federal Reserve required Bank of America and Citi to set aside $1.5 billion for compliance, and the OCC noted 78 % of banks used third-party cloud. Today, the Bank of England, FCA, and Moody’s warn that financing by Bank of America, Morgan Stanley, Blackstone and Broadcom echoes past compliance pressures. Consequently, regulators may impose stricter AI liability rules, forcing lenders to secure $63 billion deals with protective clauses like Broadcom’s residual-value support.

1 of 3

Historical Context

The 2010 Dodd-Frank Wall Street Reform and Consumer Protection Act, enacted after the 2008 financial crisis, imposed sweeping oversight on banks. The Federal Reserve required Bank of America and Citi to set aside $1.5 billion for compliance, and the OCC noted 78 % of banks used third-party cloud. Today, the Bank of England, FCA, and Moody’s warn that financing by Bank of America, Morgan Stanley, Blackstone and Broadcom echoes past compliance pressures. Consequently, regulators may impose stricter AI liability rules, forcing lenders to secure $63 billion deals with protective clauses like Broadcom’s residual-value support.

1 / 3